In typical Mobile Ad-Hoc Network (MANET) applications, the network topology changes frequently due to device movement, or link failure. Consequently, a fast response to connectivity changes is one of the most important challenges in MANETs. Most proactive routing protocols detect changes by exchanging control messages. The Neighborhood Discovery Protocol (NHDP) specifies the exchange of these kind of messages. As shown in prior studies, the proposed fixed intervals of NHDP cause unnecessary protocol overhead or late link detection. Our Adaptive HELLO (AH) scheme adapts the interval dynamically with respect to the actual situation and coordinates it between the nodes when necessary. The result is a fast link change detection and an overhead reduction in certain scenarios.

From time to time, global routing is affected by local misconfigurations of global impact. Incorrect routing information leaks into the global routing IP address space becomes hijacked. As such incidents are highly problematic and may even serve as blueprints for purposeful attacks, concepts to prevent or at least automatically detect such incidents are desirable. As planned architectures to prevent prefix hijacking incidents will need some time to become globally deployed and may be ineffective if manipulations are obfuscated, automatic detection seems to be an crucial facet. In this paper, we study a misconflguration event that could serve as blueprint for attacks. With this analysis, we try to provide starting points for the development of indicators to detect routing manipulations and classify them automatically.

Different species of malicious software (malware) have been around for quite a while. Add a command and control structure - and here you are: A "cyber army" of hijacked machines waiting for the commands of the so-called "bot herder" ready to serve the master's will. Botnets may be used for distributing spam, for installing additional malware, for information harvesting, for distributed denial of service attacks and for other actions initiated and controlled by the bot herder. Today, thousands of botnets are well understood. Their actions are observed and in some cases controlled/limited. In addition, experts active in this field argue that there is a very large number of botnets escaping tracking efforts by mechanisms such as frequent reconfiguration and frequent migration of command-and-control structures. In his keynote, Peter Martini will comment on the challenge of detecting botnets, on aggregation and clustering of similar species of malicious software and on countermeasures used today. He will comment on the relevance of botnet size and the problem of measuring the current size of well-known botnets. Finally, he will comment on legal issues and missing pieces in the fight against botnets: Botnets have come to stay.

This paper introduces a new method to predict performance requirements of mobile devices' software tasks using system models describing the hardware and software. With the help of clustering algorithms and linear regression, behavioral models of software tasks are generated automatically. These models are used to project the runtime of representative parts of the software tasks. The runtime of representative execution parts is determined with instruction-accurate simulations which are not feasible for whole executions. The inputs for the projection task a model of the hardware platform and input data parameters, especially the data size. A major advantage of this approach is that the developers do not have to estimate the performance requirements themselves. In this way the method helps to seamlessly integrate the performance analysis process into the development process. The paper introduces the ideas in detail and presents an evaluation of the proposed method for typical software tasks of mobile devices.

Abstract Topology graph based anomaly detection (TOGBAD) is a centralized approach to detect routing attacks in tactical multi‐hop networks. It uses anomaly detection based on topology graphs to identify attackers trying to launch routing attacks. Such attacks are a serious threat to multi‐hop networks since they may be used to disrupt, eavesdrop or manipulate the network. In this work, we present studies on the impact of an attacker launching a routing attack. Furthermore, we show a reasonable choice of attack parameters for the attacker. In our main contribution, we introduce TOGBAD and an evaluation of TOGBAD concerning its detection rate with and without packet loss. Finally, we illustrate TOGBAD's detection rate with an attacker trying to inure TOGBAD's anomaly detection to his attack. Copyright © 2010 John Wiley & Sons, Ltd.

Today, spreading global routing information within an Autonomous System (AS) is usually implemented by means of internal BGP (iBGP), a particular operational mode of the Border Gateway Protocol (BGP). Considering the research activities around BGP, anomalies turned out to be one major issue of iBGP: But even if the trouble iBGP anomalies induce in practice is well known for nearly a decade, new iBGP extensions still define highly problematic protocol modifications. Apparently, protocol designers are either not aware of the conceptual defects that induce anomalies or validating correctness appears to be too complex. In this paper, we provide a basis to solve both problems. Focussing the basic correctness properties of routing protocols, we will show that iBGP anomalies are caused by five root causes. These causes provide protocol designers with a tool set to reveal problematic properties of a concept and verify correctness. Based on exemplary studies for two common iBGP architectures, we will provide detailed guidance for practical usage.

Realistic modeling of the nodes' mobility is essential for obtaining credible results from the simulative performance evaluation of wireless multi-hop networks. However, most of the mobility models in the literature have not been validated with real world movement traces. To overcome this issue, we follow a trace-based approach to mobility modeling, where movement traces from the real world scenario are statistically analyzed and used for parametrization. In this paper, we introduce and evaluate a new realistic mobility model for first responder scenarios based on such an analysis while also considering geographic restrictions by incorporating free publicly available map data.

In this paper we present a new approach for identifying the crypto routines in different types of malware. In traditional malware analysis, like sandboxing, network data is examined as seen on the wire or data is collected as it is written to a file. The use of proprietary binary formats, obfuscation, or encryption hides important details, which are necessary for investigating malicious behavior. It is hardly possible to create decryptors just from monitored sandbox data. Our approach not only examines the data when leaving or entering the malware but also correlates it with information from inside the malware. By monitoring the data at I/O interfaces as well as data dependencies our approach automatically reveals the data origin. Knowing the data origin enables an analyst to easily find the crypto functions. Using this approach, we were able to identify the encryption, decryption, and command parser in different malware samples each within minutes. In our evaluation, we present the results for the Kraken command&control protocol encryption and for the file encryption of the Srvcp trojan.

Especially for large Network Operators (NOs), eliminating routing anomalies is an important aspect for the internal BGP design. To avoid such unwanted effects, classical architectures have need of certain restrictions kept in both, the network design and the iBGP peering. However, these restrictions are in conflict with the optimized network designs NOs seek. In this study we develop an inherently anomaly-free iBGP architecture that takes the demands of NOs into account. It is based on a central server that guarantees consistent local views in the entire system. This is done by exchanging additional routing information with the Border Routers of the Autonomous System. The architecture implements the results of iBGP analyses made in. Despite not being that as flexible as classical information reduction techniques, our design scales equal or better in practice. All required protocol extensions are supported by IETF Internet Drafts from persons affiliated with important router vendors.

This contribution presents validation results of an intuitive approach named `GrADAR' for automatically selecting response measures to DoS attacks. It creates and maintains a model of a computer network and of the availability of its resources from the observations of deployed monitoring systems. The graph-based model is able to express both the effects of DoS attacks and response measures as reactions to the attacks. Certain properties of the model graphs are utilized to determine different metrics which are well-known from the pragmatic decisions of network security officers.

In the recent years voice over IP (VoIP) telephony started to migrate from research to the market. In the future, all-IP networks will substitute the classical public switched telephone networks (PSTNs). Nowadays, there is no All-IP network yet, but many VoIP-providers already enable calls from VoIP to a PSTN and vice versa. By doing so, critical infrastructures within the PSTN like public safety answering points (PSAP), are accessible from the VoIP network (e.g. the Internet). Thus, there is a need for reliable performance modeling and evaluation. One aspect of particular interest e.g. for the performance evaluation of intrusion detection architectures for emergency call services is the characterization and modeling of emergency call length and frequency. In this paper, we provide a detailed analysis of traces from different PSAPs. Our work is based on empirical long-time measurements at two PSAPs. Based on these traces we characterize the load's interarrival times and call lengths concerning variation of the load, dependencies, and scalability. Furthermore, we provide fittings of the empirical data to standard probability distributions.

We present a parallel discrete-event simulation method for wireless networks such as mobile ad-hoc networks, mesh networks, or sensor networks. Our method combines new ideas with established approaches and gains considerable speedups over sequential simulations in scenarios commonly used in wireless network simulation studies. In scenarios with only 1000 nodes, the method yields superlinear speedup on up to 6 processor cores without affecting simulation accuracy. A scalability study in scenarios with 2000 nodes demonstrates a speedup of 14 on a multi-core computer with 16 cores. This is good news for a domain generally considered challenging for parallel simulation.

In the recent years Voice over IP (VoIP) telephony started to migrate from research to the market. In the future, All-IP networks will substitute the classical Public Switched Telephone Networks (PSTNs). Nowadays, there is no All-IP network yet, but many VoIP-providers already enable calls from VoIP to a PSTN and vice versa. Thus, critical infrastructures within the PSTN like the emergency call service, are accessible from the VoIP network (e.g. the Internet) and get exposed to new security threats. In particular, there is the risk of Denial of Service (DoS) attacks originating from the VoIP network. An attacker could jam the emergency call service by generating a massive load of faked emergency calls, which could lead to the loss of lives in the worst case. For us, this was the motivation to analyse the applicability of the concept of Intrusion Detection (ID) in the emergency call context and develop an adapted ID-architecture including its implementation. In an evaluation of the ID-architecture, using real emergency call traces from the fire department of Cologne, we show that the developed concept can reliably detect emerging DoS attacks from VoIP networks up to a certain VoIP diffusion rate.

Analyzing and modelling traffic is one important step in the performance evaluation of communication systems. In this paper we focus on evaluation and parameterization of traffic models for disaster area scenarios invented in our previous work. The goal is to provide scalable models for public safety scenarios as realistic and intuitive as possible. During our evaluation we examine synthetic traces for different parameters and levels of detail. As measures we examine traffic intensity, burstiness, and dependencies. The evaluation shows that the models are scalable without loosing specific characteristics. Finally, we provide an intuitive and scalable parameterization for our models that eases up using them.

Scalability is an important aspect when spreading externally learned routing information via the Border Gateway Protocol within Autonomous Systems. However, research showed that common iBGP scalability techniques like Route Reflection and AS Confederations are involved with anomalies in general. These unwanted effects can be inherently avoided by providing routers with individual sets of routing information. Implementing a modified iBGP routing information exchange in productive systems requires a scalability which is comparable to conventional techniques. Our studies show that providing routers with suitable individual information should not cause any serious decline at this point: While the most crucial parameters proved to be not problematic, only the number of paths managed by routers turn out to be hard to predict. Hence, we examine this parameter for the productive system of a large Service Provider, showing only irrelevant increases. Our results indicate that this approach could be an alternative for future iBGP routing architectures.

Simultaneous events are a fundamental challenge in distributed discrete-event simulation: Different execution orders may lead to different simulation results. Commonly, simultaneous events are handled by tie-breaking rules in order to guarantee reproducibility. A second approach, however, is examining different execution orders with a branching mechanism. By analyzing the effects of simultaneous events, confidence in simulation results may be increased. Naturally, branching may get expensive with large numbers of simultaneous events. Thus, efficiency is crucial for branching in order to be a practical method. In this paper, we present an efficient branching mechanism for simultaneous events in distributed simulation, which may be used in conjunction with, or as an alternative to tie-breaking rules. We apply techniques which optimize the detection of simultaneous events and reduce the number of branches as much as possible. Furthermore, we use cloning to share computations among different branches.

This paper deals with the characteristics of multicast push to talk voice traffic sources for disaster area scenarios. The goal is to design models that can be used to assign voice traffic sources to nodes. The modelling is based on an analysis of real-life measurements during a catastrophe maneuver. The analysis shows that about half of all calls originate from the communication head of a talk group. Based on this characteristic, different models are considered. Synthetic distributions of traffic sources for the different models are generated and evaluated by statistical analysis. Finally, the impact of the different models is evaluated in an exemplary simulative network performance analysis.

Analysing and modelling traffic is one important step in the performance evaluation of communication systems. In this paper we focus on first responder (FR) networks. The goal is to figure out models that can be used to generate realistic synthetic push to talk voice traffic for single talk groups to be used in network simulation. Our work is based on an empirical long-time measurement of one FR channel. The analysis of the trace shows significant short- and long-range correlations as well as variations of load over time. As the characteristics analysed are similar to ones observed in disaster area (DA) traces, we consider n-state and 3-state (semi)-Markov models. After fitting the parameters of these models to the traces, synthetic traffic streams for the different models are generated and finally evaluated by both visual and statistical analysis.

For demanding Grid applications that use resources at different cluster sites a reservation mechanism can ensure the availability of resources to guarantee a certain level of QoS. Besides computational and storage resources, we see the network as a reservable resource as well. Our approach to QoS is to schedule resources in the network for exclusive usage in advance. For this purpose, a flexible reservation mechanism is needed to meet the requirements of Grid applications and schedulers. In this paper we introduce a wider notion of network service than previous work. Special focus is on a file transfer service that guarantees the timely availability of data at the different Grid sites. To cope with these requirements, strategies and algorithms for a flexible network reservation service are introduced and evaluated systematically. Here, algorithms using a single path strategy are compared to algorithms using multiple paths. Our simulation results show that data transport via multiple paths leads to a higher user and network provider satisfaction than single path transfers.

Black hole attacks are a serious threat to communication in tactical MANETs. In this work we present TOGBAD a new centralised approach, using topology graphs to identify nodes attempting to create a black hole. We use well-established techniques to gain knowledge about the network topology and use this knowledge to perform plausibility checks of the routing information propagated by the nodes in the network. We consider a node generating fake routing information as malicious. Therefore, we trigger an alarm if the plausibility check fails. Furthermore, we present promising first simulation results. With our new approach, it is possible to already detect the attempt to create a black hole before the actual impact occurs.

This contribution presents a graph based approach for modelling the effects of both attacks against computer networks and response measures as reactions against the attacks. Certain properties of the model graphs are utilized to quantify different response metrics which are well-kown from the pragmatic view of network security officers. Using these metrics, it is possible to (1) quantify practically relevant properties of a response measure after its application, and (2) estimate these properties for all available response measures prior to their application. The latter case is the basis for the selection of an appropriate reaction to a given attack. Our graph-based model is similar to those used in software reliability analysis and was designed for a scalable granularity in representing properties of the network and its components to be protected. Different examples show the applicability of the model and the resulting metric values.

This paper provides a model that realistically represents the movements in a disaster area scenario. The model is based on an analysis of tactical issues of civil protection. This analysis provides characteristics influencing network performance in public safety communication networks like heterogeneous area-based movement, obstacles, and joining/leaving of nodes. As these characteristics cannot be modelled with existing mobility models, we introduce a new disaster area mobility model. To examine the impact of our more realistic modelling, we compare it to existing ones (modelling the same scenario) using different pure movement and link based metrics. The new model shows specific characteristics like heterogeneous node density. Finally, the impact of the new model is evaluated in an exemplary simulative network performance analysis. The simulations show that the new model discloses new information and has a significant impact on performance analysis.

This paper deals with voice communication models for disaster area scenarios. The goal is to design models that can be used to generate realistic push to talk traffic for single talk groups. The modelling is based on an analysis of empirical measurements during a catastrophe maneuver. The analysis shows that the time series comprise heavy load periods and significant correlations. Based on these characteristics, different Markov and semi-Markov models are considered. Synthetic traffic streams for the different models are generated and evaluated by visual and statistical analysis. Finally, a case study outlines the impact of the different traffic models in network performance simulation

In this paper we present a methodology for a performance aware design of communication systems including protocols and devices. The goal is to evaluate the performance already in early stages of the design process to avoid costly re-designs of bottlenecks in finished products. From protocol specifications given as sequence diagrams, we derive multiclass queueing networks as a means to estimate the performance of the system architecture before any executable prototype exists. This makes queueing theory accessible to system and protocol designers even if the designer is not familiar with details of queueing network theory. The methodology supports an easy evaluation of design alternatives without affecting the functional model. To achieve this, it incorporates the performance information into the system model in a non-invasive fashion, such that the system model remains meaningful without this information

Disaster areas have been figured out as a typical usage scenario for mobile wireless ad-hoc networks (MANETs). In contrast to this, there are no specific mobility or traffic models for MANETs. We present a realistic approach to realize mobility in disaster areas based on tactical issues of civil protection. The new model is analyzed and compared to Gauss-Markov and random waypoint mobility models. Furthermore, we present first simulation results. The mobility model analysis as well as the simulation are based on two real disasters that occurred in Germany in 1999 and 2001. We show that disaster area scenarios have specific characteristics. Thus, they should be considered in MANET performance evaluation.

We examine the general influence of transmission power control on a wireless multi-hop network's throughput and compare topologies resulting from different transmission power control strategies. For this purpose, we introduce concepts for measuring certain characteristics of wireless multihop network topologies.

In this paper, we introduce a simple distributed algorithm that assigns appropriate individual transmission powers to devices in a wireless ad hoc network. In contrast to many other proposed algorithms, it does not depend on special hardware. It requires only local neighbourhood information and therefore avoids flooding information throughout the network. Finally, the cooperative nature of the algorithm avoids that devices cause excessive interference by using unnecessarily high transmission powers. We show by means of simulation that the topologies created by this algorithm without any global knowledge are as effective as topologies resulting from a good choice of a common transmission power (which would require global knowledge) in terms of the achievable throughput.

This paper presents a model for the expected throughput in a Bluetooth network in presence of a number of other Bluetooth networks that cause radio interference. The analysis considers unidirectional traffic and makes worst-case assumptions on the effect of radio interference. In order to verify the model, simulation results are presented. The results from this work may be very useful when designing ad-hoc networking functionality for future versions of Bluetooth.

In this paper, we introduce statistical methods to estimate the stability of paths in a mobile wireless ad hoc environment. Identifying stable paths helps to reduce control traffic and the number of connection interruptions. By means of simulation, we analyse the stability of paths chosen according to a variety of strategies, including those used by the well-known routing protocols, AODV and DSR, under a variety of different mobility patterns. This offers new insights into the relation between a path's stability and other characteristics and shows that our statistical metrics are able to identify stable paths in a wide range of scenarios.

The vision of ad-hoc networking with Bluetooth includes the concept of devices participating in multiple "piconets" and thereby forming a "scatternet". However, the details of scatternet support for Bluetooth are not specified yet. This paper presents a scheme for Bluetooth scatternet operation that adapts to varying traffic patterns. Being based on sniff mode, it does not require substantial modification of the current Bluetooth specification and may thus be incorporated into currently available Bluetooth products. We present simulation results that confirm the applicability of our approach to realistic scenarios.

We develop adaptive metrics to identify stable links in a mobile wireless networking environment based on the analysis of link durations in several different mobility scenarios. Our metrics only rely on online statistical evaluation of observed link durations. Neither do they require information on signal strength, radio conditions, or spacing of the mobile devices, nor do they depend on the availability of additional hardware such as GPS receivers or a synchronisation of the devices. We demonstrate the ability of the metrics to select stable links with a high probability in a wide range of scenarios.

Differentiated Services is a very promising approach to achieve quality of service in the Internet. However, setting up static service level agreements (SLA) between different Internet service provider domains lacks flexibility and may reduce quality of service as resource reservations cannot be adapted to the actual traffic. With the introduction of bandwidth brokers, a significant step towards a completely dynamic Differentiated Services environment has been made. Bandwidth brokers negotiate and automatically set up SLA according to network load and provider policies. Nevertheless, this automatic set-up may fail if there are bottlenecks in the network caused by physical limitations or policy issues and may lead to congestion. In this paper, we propose a scalable mechanism to handle these congestion situations, that uses four phases each solving a different level of congestion. Our mechanism does not need any destination information thus making it applicable for the majority of Internet applications.

Future applications of Bluetooth are likely to include ad-hoc networking. Therefore, it is desirable to interconnect multiple Bluetooth piconets to form a scatternet. Up to now, there is no extensive proposal for scatternet support available. We present and analyze an adaptive scheme for scatternet scheduling that is based on the sniff mode and thus does not require substantial modification of the current Bluetooth specification. The suitability of our approach is shown by first simulation results.

The migration of the Internet from classic computer communication to a platform for multimedia applications with real-time communication requires end-to-end improvements of the network-level service. The proposal known as Differentiated Services is a very promising approach for implementing quality of service in the Internet and is being discussed and developed. The challenge is to deliver end-to-end QoS on top of Differentiated Services considering multiple concurrent application level data streams. We propose a QoS management system for multimedia servers that benefits from the scaling properties of layered media streams. This enables the system to map application QoS demands to available network resources and to adapt the quality of individual streams according to inter-stream QoS dependencies.

The BLUEPAC (BLUEtooth Public ACcess) concepts present ideas for enabling mobile Bluetooth devices to access local area networks in public areas, such as airports, train stations and supermarkets. The proposed protocols support mobility on OSI layer 3. This paper concentrates on the necessary layer 2 protocol concepts for supporting mobility and handoffs between different access points. Furthermore, we present the necessary adaptations for allowing IP datagrams to be exchanged between the mobile Bluetooth devices and access points. The BLUEPAC protocol concepts have been implemented above a Bluetooth emulation system to test their feasibility. First results clearly showed the importance of minimizing the handoff duration to allow efficient operation of upper layer protocols such as TCP.

We present protocol concepts for an extension of IP for mobility issues in Bluetooth networks. The protocol is called BLUEPAC IP, where BLUEPAC stands for "BLUEtooth Public ACcess". "Public access" means access to various kinds of information in public areas, e.g. airplane, train, hotel room, department store, museum. Bluetooth will reside in low-cost chips that provide wireless communication up to OSI Layer 2. By the end of this year many mobile devices, e.g. notebooks, PDAs, cell phones, will contain Bluetooth chips as a cable replacement. But Bluetooth is move powerful: with a suitable network protocol that inherits all aspects of IP, it will be possible to provide IP services over Bluetooth. BLUEPAC IP takes IP as a basis and additionally includes functionalities of mobile IP and cellular IP for local IP address assignment, routing issues and handoff support.

Previous articles by Frank and Martini (see 22nd Annual Conference on Local Computer Networks, LCN '97, Minneapolis, 1997 and 23rd Annual Conference on Local Computer Networks, LCN'98, Boston, 1998) on the concept of ROBIN (Regulation Of Bandwidth in Intra-Networks) gave a motivation for and presented first results of our approach to regulate the sending bandwidth of data flows in heterogeneous campus networks by determining bottlenecks and possible rate limitations with a global knowledge of the network state based on the end-to-end operation of the mechanism. This paper presents the latest results of our performance study of the ROBIN approach with a focus on the scalability and the concept of data flow aggregation within the bandwidth regulation. The QoS properties of ROBIN together with the flow aggregation for the local domain show a close relation to the ideas of the differentiated services activities for QoS support in the Internet, which is also discussed in this paper.

The variety and heterogeneity of different approaches for Quality of Service (QoS) support on different protocol layers result in the need of a solution with end-to-end overview and regulation of dataflows in a network. Previously we proposed the concept of ROBIN (regulation of bandwidth in intra-networks) to regulate the sending bandwidth of data flows in heterogeneous campus networks by determining bottlenecks and possible rate limitations with a global knowledge of the network state based on the end-to-end operation of the mechanism. This paper presents latest results of a simulation-based performance analysis studying the throughput fairness of an enhanced version of ROBIN and evaluating the impact of the bandwidth regulation on delays and packet losses in the network.

The inclusion of integrated quality of service (QoS) support in existing and future networking technologies and protocols has become a highly relevant area of research. But the variety and heterogeneity of different approaches for QoS support on different protocol layers result in the need of a solution with an end-to-end overview and regulation of data flows in a controllable subnetwork. The ROBIN (Regulation Of Bandwidth in Intra-Networks) concept introduced in this paper is such a mechanism which controls the bandwidth used by data flows in heterogeneous networks in case of network congestion. This paper presents an overview of the concept and reports on the state of simulative performance analysis work on ROBIN.

The paper presents an extension of the transport layer to control the bandwidth used by data flows in heterogeneous networks in case of network congestion. The regulation mechanism is operating on an end to end basis. It does not depend on a negotiation of quality of service (QoS) parameters prior to the data transmission. Instead it uses a relative QoS concept based on a classification of data flow requirements specified by the application. The paper describes the concept of the mechanism for Regulation Of Bandwidth in Intra-Networks (ROBIN) and presents first measurement results with a prototype implementation of ROBIN.

The demand-priority protocol currently in the process of standardization by IEEE 802.12 aims at supporting interactive multimedia applications by providing a low latency service for high-priority traffic. This goal may be achieved in case of large frames and/or small distances. Otherwise, improvements are required. This paper starts with a brief description of the basic characteristics of the network topology and the medium access control protocol. Then, it presents simulation results for normal and high priority traffic in scenarios with variable bit rate high priority loads for networks of different sizes. It reveals the drawbacks of the demand-priority protocol and clearly shows the advantages of using service strategies different from those included in the current version of the draft standard.

We study the performance of the demand-priority LAN currently under standardization by the IEEE LAN/MAN Standards Committee. The corresponding working group named IEEE 802.12 has already finished the standard of the base protocol which does not yet meet all goals stated for this new high speed LAN. The paper shows that the protocol is highly sensitive to traffic characteristics of the high priority load. In particular, it suffers from poor efficiency in case of small high priority frames.

We study the throughput characteristics of a FDDI workstation cluster installed at the University of Paderborn. We compare theoretical results to the performance really available to the user at the service access point to TCP and UDP. Furthermore, we compare the performance for two difficult operating systems: we consider the SUN-OS based on BSD and the new Solaris 2.3 developed for multiprocessor environments. The measurement results showed a maximum throughput of less than 60 Mbps which is still considerably less than the bandwidth promised by FDDI vendors. Moreover, it becomes clear that high throughput absolutely requires the transmission of large frames. We also briefly address the current discussion on higher speed LANs within IEEE 802.>

The interworking modes for the interconnection of metropolitan area networks (MANs) and asynchronous transfer mode (ATM)-based broadband-integrated services digital network (ISDN) are studied. The modes considered are "cell-to-slot interworking" and "frame interworking." The cell-to-slot mode modifies ATM cells to distributed queue dual bus (DQDB) slots and vice versa. The frame mode reassembles cells/slots and modifies AAL frames to DQDB IMPDUs and vice versa. Cell/segment delays for both modes in an interworking unit (IWU) receiving aggregated traffic from B-ISDN and using high priority transmission controlled by the guaranteed bandwidth (GBW) protocol for access to DQDB are compared. The performance results show the advantages of the cell-to-slot mode.

The authors study packet loss in high speed networks interconnecting conventional local area networks (LANs). They consider two different techniques to control contention at the station where packets leave the high speed network and must wait for service by the LAN. One technique is based on a backoff strategy where packet transmission to a temporary overloaded station is interrupted for a randomly chosen time. The other technique studied uses rate-based flow control. A discussion of results for both fiber distributed data interface (FDDI) and distributed queue dual bus (DQDB) shows protocol-specific peculiarities requiring different solutions to different problems. Packet loss characteristics obtained for FDDI and DQDB are very close to each other.>

Two strategies which have recently been proposed for bandwidth allocation in distributed-queue dual-bus (DQDB) subnetworks of a metropolitan area network (MAN) are compared. Studied are packet delays for connection oriented variable-bit-rate (VBR) traffic such as real-time data and compressed voice and video. It turns out that the guaranteed bandwidth protocol is an excellent candidate for connection oriented data service in DQDB where support of delay-sensitive VBR traffic is a major concern. In contrast to this, a modified version of the bandwidth balancing mechanism suffers from slow convergence to the desired bandwidth allocation.>

The pros and cons are discussed of synchronous and isochronous service in FDDI-II, which is an enhanced version of the Fiber Distributed Data Interface (FDDI) protocol. The results make clear that basic FDDI meets the requirements imposed by traditional LAN applications. To overcome the problem of throughput degradation and increasing delays for asynchronous traffic in applications with large fiber length and hard response time requirements, FDDI-II is defined as a promising upward-compatible extension. It is shown that from a performance point of view FDDI-II is an excellent candidate for complex communications environments. This protocol integrates both circuit switched and packet switched services in an efficient way. The results show that, in environments with upcoming hard real-time requirements, changing to the enhanced protocol, i.e., changing from basic FDDI to FDDI-II, is a much better way to meet the demands than to push the old protocol to its limits.>

Recently, a discussion came up dealing with the frame copied indicator in the frame status sequence of FDDI. This indicator shows whether or not the destination has copied the frame and this way allows the sender to react in a proper way. However, in interconnected networks the destination node may represent a large number of destinations with different addresses making it difficult to check within a very short time whether a certain packet is to be copied and the indicator is to be set. The paper discusses the effect of using or not using the frame copied indicator. Additionally, it proposes a class of new congestion control strategies for FDDI.>

The integration of group 4 facsimile in local area networks (LANs) based on carrier-sense multiple-access with collision detection (CSMA/CD) is studied. A simulation program yields performance characteristics such as message and packet transmission delays, channel access delays and average throughput. Traffic on the LAN is modeled by group 4 fax generators and so-called base load generators. Both classical Poisson processes and interrupted Poisson processes are studied. A comparison of the simulation results shows the impact of the kind of base load generation on facsimile service. The sensitivity to the intensity of the generated facsimile and base load and to flow control mechanisms implemented on logical link control (LLC) or transport layer is evaluated.>

Performance issues of the FDDI-II protocol, an upward-compatible enhancement of the basic fiber distributed data interface (FDDI), are discussed. The main characteristics of the FDDI protocol are examined. The performance of FDDI-II is studied for scenarios including asynchronous traffic, synchronous traffic and isochronous traffic. The results make clear that basic FDDI meets the requirements imposed by traditional local area network (LAN) application and that from a performance point of view FDDI-II is an excellent candidate for complex communications environments. When compared to the distributed-queue-dual bus (DQDB) protocol, FDDI-II yields much better efficiency in traditional LAN applications.>

Packet delays imposed by state-of-the-art bridges are compared to delays imposed by the high-speed local area network (HSLAN) itself. The communication system for three different types of internet traffic, namely file transfer, interactive traffic, and a mixture of both is discussed. The analysis shows that even for long transmission medium, service capabilities of the HSLAN will not be the bottleneck but that, bridges available today generally are not able to process packets as fast as they can be forwarded by the HSLAN.>

The draft distributed queue dual bus (DQDB) protocol is a proposed standard for metropolitan area networks (MANs). A study is made of fairness issues of the current version of this IEEE 802.6 working document, which is intended to be the basis for MAN service offered by public network providers. Performance is studied with respect to both delays and throughput. It is shown that the throughput provided to specific users may depend on the geographic location with respect to other users. On the other hand, DQDB yields fair medium access for the majority of realistic scenarios.>

The authors study multiple-token rings for mixed traffic imposed by different applications and aiming at the attainment of high-speed local area networks (HSLANs). An approximate analysis based on the first two moments of interarrival time and token rotation time yields results extremely close to simulation results. The multiple-token ring appears to be an excellent candidate for the envisaged scenario. However, buffer dimensioning at the destination LAN, studied by simulation only, proves to be one of the most critical factors during the design phase of a HSLAN.>

A model of a multiple-token-ring backbone based on fiber-optic transmission, intended for LAN interconnection, is presented. Because it is one of the most important applications for inter-LAN communication, file transfer is selected for performance evaluation. A mixed traffic environment including interactive traffic is assumed. The suitability of multiple-token-access protocols is studied by approximate analysis and simulation, yielding results for both queue length and packet delay. Sensitivity with respect to medium length and station latency is considered. The results allow performance predictions for high-speed backbone networks.>